InsightIQ

01 / The Challenge

A dashboard nobody wanted to open

InsightIQ had all the data marketing teams needed, but it was buried across six separate views with no connective tissue between them. Users had to click through an average of eight screens just to compare two metrics side by side. The result was predictable: only 34% of users bothered to check the dashboard on a weekly basis.

The business impact was real. Customer success teams fielded dozens of tickets each week from users who couldn't find the numbers they needed. Churn among mid-market accounts had crept up to 18% quarterly, and exit surveys consistently pointed to a "confusing" and "overwhelming" product experience as the top reason for leaving.

02 / Research & Discovery

Listening to power users and studying the landscape

I kicked off the project by scheduling 12 in-depth interviews with InsightIQ power users across different company sizes and industries. These were the people who lived in the product daily, and their frustrations painted a clear picture: the dashboard felt like a filing cabinet where every drawer was labeled in a different language.

Alongside interviews, I ran a competitive audit of Amplitude, Mixpanel, and Datadog, focusing on how each product handled data density, customization, and first-time user experience. The standout insight was that the best dashboards gave users control over their own layout without requiring them to build everything from scratch.

• 12 stakeholder interviews with marketing leads, data analysts, VPs, and growth PMs across 8 companies
• Competitive audit of Amplitude, Mixpanel, and Datadog focusing on dashboard patterns, widget systems, and onboarding flows
• Identified 4 key personas: Marketing Lead (needs quick KPIs), Data Analyst (needs deep drill-downs), VP Marketing (needs shareable executive views), and Growth PM (needs experiment tracking)
• Synthesized interview notes into an affinity map that revealed 5 recurring themes: data fragmentation, alert fatigue, lack of customization, poor mobile experience, and unclear data hierarchy

03 / Information Architecture

Making 47 data points feel like 5

The old dashboard tried to show everything at once, which meant nothing stood out. I needed to restructure the information architecture so the most important metrics surfaced immediately while deeper data remained accessible without cluttering the view.

I ran a card sorting exercise with 8 participants, asking them to organize all 47 data points the platform tracked into groups that made sense to them. The results were remarkably consistent: users naturally clustered data into five logical categories: Performance Overview, Channel Breakdown, Campaign Health, Audience Insights, and Custom Reports.

This structure became the backbone of the new widget system. Instead of fixed pages, users would see a default dashboard built from these five clusters, with the ability to add, remove, reorder, and resize individual widgets. The idea was to give everyone a great starting point while letting power users make it their own.

• Card sorting with 8 participants across all 4 personas
• 47 data points reorganized into 5 logical clusters
• Designed a customizable widget system with smart defaults
• Created progressive disclosure patterns so summary cards expanded into detailed views on click
• Built a "Quick Compare" feature letting users drag any two metrics side by side

04 / Design Exploration

Three concepts, one clear winner

With the information architecture locked, I explored three distinct dashboard concepts in wireframe form. The first was a fixed layout with a rigid grid of cards, similar to Google Analytics. The second was a fully modular widget system where users could drag and drop anything anywhere. The third was a guided insights approach that used AI to surface relevant metrics based on the user's recent activity.

I tested all three concepts with 6 users in moderated sessions using clickable Figma prototypes. The modular widget approach won decisively, but with an important caveat: users wanted smart defaults, not a blank canvas. Nobody wanted to spend 20 minutes setting up their dashboard before they could use it. The final direction combined the flexibility of the widget system with curated templates based on the user's role.

05 / Visual Design & Iteration

Dark theme, dense data, clear hierarchy

The visual design needed to handle serious data density without overwhelming users. I chose a dark theme as the default because it reduces eye strain during long analysis sessions and makes colored data points pop against the background. Teal served as the accent for positive metrics and growth indicators, while a warm red flagged alerts and declining trends.

I built 32 chart components in Figma, each designed to work at three different widget sizes (small, medium, and large). Every chart had consistent interaction patterns: hover for detail, click to expand, and right-click for export options. Typography followed a strict hierarchy with numbers rendered in a tabular font so columns always aligned cleanly.

The iteration process was relentless. I ran two additional rounds of usability testing on the high-fidelity prototype, catching issues like insufficient contrast on secondary labels and confusing iconography on the widget picker. Each round tightened the design until every element earned its place on screen.

06 / Outcomes

From dreaded to daily

The redesigned dashboard launched to all InsightIQ customers over a phased two-week rollout. The impact showed up in the data almost immediately. Weekly engagement more than doubled, jumping from 34% to 71% within the first month. Users were spending 2.3x longer in the product per session, and the average time-to-insight dropped by 40%.

NPS climbed from 23 to 52, and the customer success team reported a significant drop in dashboard-related support tickets. Most telling was the qualitative feedback: users who had previously built workarounds in spreadsheets started sharing InsightIQ dashboards directly in their team Slack channels.